Hi everyone,
I am hitting a brick wall with the Local_Info_Agent simulation in VS Code. Despite following the "manual permission assignment" fix mentioned by Eduardo and Swati and by Bhoopesh, I am still receiving the generic "Something went wrong. Internal Error, try again later" message every time I click "Start Simulation."
Here is everything I have verified and executed so far:
- Org Setup: Confirmed "Turn on Einstein," "Einstein Bots," and "Agentforce (Default)" are all toggled ON in the Developer Edition org.
- User Configuration: Verified the default_agent_user in the .agent file matches the generated afdx-agent username in the org.
- Permissions (Admin): Manually assigned AFDX_User_Perms, Prompt Template User, and Agentforce Default Admin to my Admin user.
- Permissions (Agent User): Manually assigned AFDX_Agent_Perms, Coral Cloud Admin, Coral Cloud Agent, and all standard Agentforce Service Agent permission sets to the agent user.
- Code Integrity: Verified Agent Script syntax is correct (headers like system:, config:, etc., are properly spelled and indented).
- Deployment: Successfully ran sf project deploy start to sync the metadata bundle to the org.
- Environment: Reloaded the VS Code window and validated the agent script (Validation returns "Success").
Despite a 100% successful environment setup and manual permission overrides, the simulation server still fails. Is there a known issue with the LLM gateway for this specific module, or a hidden permission (perhaps related to the Einstein Agent profile itself) that the setup script is failing to bridge?
Any insights would be greatly appreciated!
